Cardioid Bass

[Pano]

I know you weren't asking me

Well maybe not - but I'm quite happy to read your comments anyway. (and others)

I've always thought the "you can't localize bass" thing to be silly. I sure can.
We had a midsize earthquake last year. I could point right to where THAT super LF came from.
Same for bass cabinets when I used to run concert PA.

But that was all outdoors, or in very big rooms. Earl is the first one I've read to say you can't localize it in a small room. Hadn't thought about that before. Where would the localization cutoff be? The room fundamental?

 

[Graham Maynard]

Hi TerryJ,

You describe exactly what I was writing about relating to arrival time.
If our physical feeling does not match our hearing, then reproduction does not appear real - no matter what the measurements say.
You managed to obtain coincidence for first wave at the listening position, but what about the timing of all the other energies within the room and differently audible at other positions ?

Hi Panomaniac,

As far as I am aware it is possible to level/tune a simple box LS placed behind a baffle to approximate LF cardioid, and its tuning could be made to counter the most significant listening room weaknesses without the overall response needing to be perfect cardioid at all. JohnK has recently proposed that pure cardioid does not necessarily provide best room output.

Hi Rybaudio,

A flat frequency response does not guarantee a flat phase response, and if the 'waves' are not radiated in phase then they simply cannot be heard 'in time'.
It is easy to amplitude compensate for tuned enclosures, driver Q etc. and thereby obtain a flat frequency response; however, to ensure that radiation is in-phase and coherent is an entirely different matter.
The difference might not be recognisable on home theatre, but it sure is on an orchestral kettle drum, a 'beat' of bowed chellos, anything recorded open air or within the ambience characterisations a large hall.

Hi K Strain,

I do so agree with your Post#120.
Another aspect being that ported subs lose out on first cycle amplitude as a portion of waveform energy is not turned into room air motion, but becomes stored within the mechanism which provides that resonant boost. Dynamics become sacrificed for SPL generation - and as you say - occasionally that stored energy is released at a moment when the waveform does not warrant it, inducing cognitively recognisable reproduction distortion.

Hi Earl,

Ever heard a very high quality car audio system; not the SPL headbangers, but those who go in for SQ competitions with all the test gear etc.?
They can select/equalise and select power amps/drivers as necessary within the confines of vehicular space and acheive genuinely superb and coherent reproduction to below the limits of our hearing. (Granted that is generally for front or between front seats listening position.)
They compensate for the change to 'listening room' pressurisation instead of wave propagation, though they don't have to think about neighbours.

A small room might not support conventional ideas of long wavelength propagation, but it can still be pressurised to provide decent reproduction, which is what you too appear to be doing with subs anyway!

Hi JohnK,

I note you have cardioid for say 50Hz down, but is this really of benefit for room listening ? (Earl's challenge.)
I feel that the most disturbing aspect for dipole response is the first corner/wall rear reflection which becomes additive to front output at about 100Hz. If this could be nulled by a cardioid response then that peak would not arise, but this would need to be via your type of resistive cardioid and then tunable for reflection distance within a user's home.
Any merit in this idea worth examining ? (I can't try these things.)

Cheers ........ Graham.

 

[MaVo]

A small room might not support conventional ideas of long wavelength propagation, but it can still be pressurised to provide decent reproduction, which is what you too appear to be doing with subs anyway!

If that would be the case, the multiple sub setup would not be necessary. They are used to counter the modal problems, which by definition occur above the room fundamental / pressure mode. A room would need to be smaller than 2,15meters in every dimension to act in pressure mode to 80hz. 3,44meters for 50hz. That works in cars, but not in listening rooms, which will most likely be bigger than this.

 

[Pan]

A flat frequency response does not guarantee a flat phase response,

Are you thinking on the effect of crossovers on a multiway loudspeaker or are you suggesting a loudpeaker driver is not a minimum phase device?

and if the 'waves' are not radiated in phase then they simply cannot be heard 'in time'.

If the groupdelay is below the threshold of the ear then the signal will most likely be heard as "in time". Or?

It is easy to amplitude compensate for tuned enclosures, driver Q etc. and thereby obtain a flat frequency response; however, to ensure that radiation is in-phase and coherent is an entirely different matter.

How can it be if a loudspeaker is a minimum phase device?

The difference might not be recognisable on home theatre, but it sure is on an orchestral kettle drum, a 'beat' of bowed chellos, anything recorded open air or within the ambience characterisations a large hall.

Yes, excessive groupdelay is shown to be audible and the ear is slightly more sensitive to this in the 100-250Hz area as I understand it. Excessive groupdelay may (for example) be the result from a subwoofer with sharp rolloff on both ends and narrow bandwith, no matter the type.

Dynamics become sacrificed for SPL generation - and as you say - occasionally that stored energy is released at a moment when the waveform does not warrant it, inducing cognitively recognisable reproduction distortion.

I'm curious how you came to that conclusion and what you think is an aceptable level of groupdelay at various frequencies?

Cheers!



/Peter

 

[john k...]

Hi TerryJ,

Hi JohnK,

I note you have cardioid for say 50Hz down, but is this really of benefit for room listening ? (Earl's challenge.)
I feel that the most disturbing aspect for dipole response is the first corner/wall rear reflection which becomes additive to front output at about 100Hz. If this could be nulled by a cardioid response then that peak would not arise, but this would need to be via your type of resistive cardioid and then tunable for reflection distance within a user's home.
Any merit in this idea worth examining ? (I can't try these things.)

Cheers ........ Graham.

First let me say that I think part of the disagreement in this thread arose because of the difference between talking about full range loudspeakers placed conventionally in a listening room and bass augmentation using subwoofer. My comments about cardioid bass are largely in respect to the problems associated with placement of full range loudspeakers in which case cardioid bass seems to have the best solution in regards to sensitivity to room/listener placement compared to dipoles or monopoles. This is a different problem than where to place a 50 Hz and lower subwoofer for best in room response. So let's realize there are apples and oranges here.

Second. I don't like the addition of subwoofer for music systems. The problem is the introduction of additional group delay, even with the woofers placed close to the main speakers, due to the speaker/sub crossover. I have discussed this at my web site. See the bottom of the page. There is, according to Martin Colloms, evidence that this additional GD is audible.

With regard to the NaO systems it is a little more complicated than it appears. If you look at the way dipoles, cardioids and monopoles
blend through the crossover region you may begin to understand the design of the NaO II and NaO Mini a little better. With the Mini the panel crosses over to a monopole woofer with an LR4 crossover at 125 Hz. The result is that above the crossover point the radiation is dipole. But at the crossover point the summed woofer + mid response moves towards cardioid thus avoiding the rear wall reflection and resulting coloration of the direct sound common to full range dipoles and monopoles in that frequency range. Below the crossover point the response warps from cardioid to monopole. The Mini plans include an Excel worksheet which assists the builder in tuning the active woofer circuit in the Mini crossover so that the woofer system has a 2nd order, Q = 0.5 response with Fs at the builder's listening room's fundamental frequency so that in smaller rooms the woofer will not over load the room below the room fundamental.

With the NaO II, you are correct in that the U-frame woofer tends to a monopole at higher frequency. That can not be avoided with an acoustic resistance enclosure. So rather than fight it I decided to take advantage of it. Again, you must consider how the woofer interacts with the main panel. Here, when the woofer is operated in U-frame format the over all response is dipole with a transition to cardioid (monopole +dipole) in the crossover region and then the natural transition of the u-frame to quasi-cardioid at lower frequencies. When the NaO II woofer is operated in monopole format the system is designed to behave in a similar manor to the Mini. I’ll be the first to admit that the execution of these design concepts may not be perfect, but they were the targets.

 

[gedlee]

BTW, do you like BP boxes because of the high SPL in the limited band you're using?
On a differnt note. What do you think of Graham's objections to LF arriving at different times or out of phase. May not be a problem with the LFE track, but how about for music?

A little late, busy last night.

I use BP subs for thier increased efficiency AND their acoustic LPF. This means that any nonlinear distortion will be filtered out acoustically. Nonlinearity in a sub, while not really a problem audibly can help you to localize the sub. Port noises are another problem. My bandpass subs all have big ports and/ or foam to quite them down.

The point about localization being audible but not necessarily phase or group delay is right on. I have heard this effect too, but I would not jump to the conclusion that it was phase effects. In fact given the evidence that is highly unlikely.

In my expereince if you can localize on any of the subs then something is wrong somewhere. It should not be possible.

In a small room in the modal region, the SPL rises rather slowly at the individual modes as the energy builds up. The point about this NOT being like a propagating sound wave is quite correct. The pressure throughout the room basically rise and falls virtually in unison at every point. Think about the rise time of a mode compared with the wave propagation time. If you plot out the pressure around the room you will see that the wave travels around faster than the pressure can respond and the entire room pressure rises and falls with no apparent wave propagtion. Once we reach the modaly dense region then this effect goes away because the wavelengths are shorter and the modes are denser.

I worked in car audio for 20 years so I know a little bit about it. There is an effect of small rooms that when the bass is smooth it can actually have a more engulfing effect than a large room. Its just that it also easy to create a small room that has a peaked and boomy response. The only small rooms that I have heard that have this effect have multiple subs.

As to localization, if you setup the subs as I have suggested then the main LF energy is coming from the mains and the other subs simple augment this. In that case the localization should always be at the mains. Again, if you can locaize on any one of the subs then its setup wrong. You CAN localize on a sub if you get very close to it, which is why I try and put them fairly far from the listener.

If you don't have a spectrum analyzer then there is a procedure that I use that "kind of works", but will definately not work as well as real measurements.

Setup the mains for maximum imaging. Add the biggest sub to blend well with the mains at the seating location. Now go very close to each other sub and set its output so that it is just audible. When you get back to the seating loaction no sub should be localizable and the response should be fairly smooth.

 

[Etienne88]

Interesting discussion!

Earl, I would like to know under which frequency you are using your multiple subs set up?
What would then be the highest frequency you recommend for the multiple subs set up? I guess somewhere below 100-200Hz since it is the region where distinct modal behavior takes place… I also have the feeling that it is room dependent, isn’t it?

Regards,
Etienne

 

[Graham Maynard]

Hi JohnK,

Thanks for your reply. I mentioned that I did not like separate subs - you state the same reason;- group delay = time/phase reproduction error wrt the original electrical waveform.

That you already have the first reflection peak covered within the crossover/*pole transition is better than I imagined.
I don't think I found this information on your Website pages.

That reflection peak is unavoidable with plain *pole loudspeakers, and can be a serious impairment. It can be EQed out, though not without weakening the dynamic response.


Hi Pan,

Thanks for your reply. Many challenges.

I don't regard either crossovers or drivers as being minimum phase devices.

Or ?
Group delay is not an amplitude effect having a threshold. It affects relationships between different frequency groups and harmonics, thus upon the 'characteristic sounds' of reproduced musical instruments.
Often this is recognisable as the 'tone' imparted upon musical indentities by any LS.


Hi Mavo,

And down at 30Hz.......
No wonder the middle region of my lounge ceilings cracked into plasterboard sized sections !


Hi Earl,

All understood !
(Maybe like a supertweeter, where if you notice it alone through the music, then it is not set up correctly.)


Cheers ......... Graham.

 

[gedlee]

Interesting discussion!

Earl, I would like to know under which frequency you are using your multiple subs set up?
What would then be the highest frequency you recommend for the multiple subs set up? I guess somewhere below 100-200Hz since it is the region where distinct modal behavior takes place… I also have the feeling that it is room dependent, isn’t it?

Regards,
Etienne

Of course its room dependent. Below about 150 Hz. for most rooms that I have dealt with, but these may be on the largish size for Europe.

 

[Etienne88]

Thanks for the reply.
By Europe, you mean the difference between timber and concrete houses, isn't it?
In Scandinavia, most of the houses are built using timber, so is mine!

/Etienne

 

[Pan]

Hi Pan,

Thanks for your reply. Many challenges.

I don't regard either crossovers or drivers as being minimum phase devices.

Or ?
Group delay is not an amplitude effect having a threshold. It affects relationships between different frequency groups and harmonics, thus upon the 'characteristic sounds' of reproduced musical instruments.
Often this is recognisable as the 'tone' imparted upon musical indentities by any LS.
Cheers ......... Graham. [/B]

Yes, but if the groupdelay is small enough it will not be audible.. therfore I used the word "threshold".

If a driver is not essentially a min phase device.. how would you describe it? I think the physics behind this is pretty well sorted out.


/Peter

 

[gedlee]

Thanks for the reply.
By Europe, you mean the difference between timber and concrete houses, isn't it?
In Scandinavia, most of the houses are built using timber, so is mine!

/Etienne

No actually I meant the physical size. European rooms tend to be smaller as far as I remember.

 

[Graham Maynard]

Hi Pan,

Drivers sorted ? If only !

None is fullrange and without need for augmentation via crossovers, and all generate transduction phase change across their own individual working ranges.

How would I describe a driver ?
Highly complex, and not only electrically.

Cheers ......... Graham.

 

[john k...]

Have to agree here Grahm. Dynamic drivers (woofer,tweeter, etc) are minimum phase. I have neven measured one which wasn't. And crossovers and filters, other than allpass filters are also pretty much minimum phase. Not that this is relevant to this thread.

 

[Graham Maynard]

AHA, Minimum phase -

I was discussing transduction phase and stored energies etc., for that is where the real-world 'hands-on' problems lie.

Cheers ....... Graham.

 

[kstrain]

Question and comments

AHA, Minimum phase -

I was discussing transduction phase and stored energies etc., for that is where the real-world 'hands-on' problems lie.

Cheers ....... Graham.

Could you please define "transduction phase"? Unless I'm missing something profound, it seems to be extraneous to the conversation.

As I'm sure you know, if a linear system is minimum phase, the time and frequency responses are uniquely related. Therefore the idea of "stored energy" is saying nothing that is additional to the frequency response.

At low frequency, and measured in the near field, the response of a speaker (box and all) must be minimum phase. That does not extend to the whole room, unless acoustically small, as where there are nodes there must be more than one solution for the phase at amplitude-wise equivalent places.

Ken

 

[gedlee]

Re: Question and comments

At low frequency, and measured in the near field, the response of a speaker (box and all) must be minimum phase. That does not extend to the whole room, unless acoustically small, as where there are nodes there must be more than one solution for the phase at amplitude-wise equivalent places.

Ken

I suppose you should be a little careful here because what level of non-minimum-phase are we talking about. For a LF signal to be detectably non-minimum phase the delay time would have to be pretty long. But there could be some small delay, which, in principle, would be non-minimum phase, but the group delay would be exceedingly small. In these kinds of discussions we have to be careful to put some bounds on things. 1 ms of group delay at 50 Hz is not a lot of phase shift - about 15°. Most research has shown that detection of group delay falls pretty fast below about 1 kHz. But then it can rise again at very low frequencies, but more related to phase shift than group delay. It's a complex subject and highly prone is complicated arguments that are completely semantic.

 

[Graham Maynard]

Hi K Strain.

I am writing about the variation of phase with frequency of the waveform as measured by a microphone (say on axis) and compared to the original electrical waveform.
This being something fundamental, most certainly not extraneous, and directly relating to the stability of a generated cardioid response with frequency.
Stored energy being part of the steady sine facet of frequency response, and meaning that cardioid may not fully hold for transients when this has been set up via sine observation.
Thus cardioid developed using low Q drivers on OBs and/or in non-resonant enclosures will hold better for music reproduction, than those not.

I am sure this can be appreciated.

Cheers.......... Graham.

 

[kstrain]

Hi K Strain.

I am writing about the variation of phase with frequency of the waveform as measured by a microphone (say on axis) and compared to the original electrical waveform.

Cheers.......... Graham.

Forgive me, but I'm still not exactly clear what you mean. I hope the microphone side is calibrated/compensated to measure sound pressure without phase shift over frequency. (I.e. distinguish the measurement system from the device under test.)

If so the important phases are those due to the low- frequency resonance of the driver, measured as cone motion (certainly minimum phase); the line of sight propagation of the sound to the microphone (pure time delay) and whatever happens to the rear and reflected sound (not necessarily minimum phase depending on the measurement setup).

Is the first of these the "transduction phase"?

Ken

 

[Graham Maynard]

Hi Ken,

Yes, and similar crossover induced transduction phase changes.

I don't like the theoretical term 'minimum phase'.
Its as bad as saying crossover components have a linear response, when actually they can directly cause serious waveform distortion - considerable difference between an original music waveform and the eventual air displacement.

Cheers ....... Graham.